In a communication system, it is desired that a bit error rate can be reduced as much as possible in a data transmission process. For example, a forward error correction (Forward error Correction, referred to as FEC) algorithm is adopted. The FEC is a technology for transmitting redundancy and transmitted data together by using a preset algorithm. A receiving apparatus has a capability of detecting and correcting, according to the redundancy, a plurality of error bits that may occur during transmission. However, because some fixed error modes exist, when an input bit error rate of a system is reduced, an output bit error rate of the system still keeps unchanged. The case that the input bit error rate is reduced while the output bit error rate is almost unchanged is referred to as an error floor (error floor), which case the system cannot tolerate.
In order to resist the error floor, in the prior art, solutions such as extracting soft decoding information for reference for a next level, increasing error correction accuracy, increasing one-time decoding accuracy, or limiting capacity in a decoding preliminary phase to avoid error reproduction exist. However, on the whole, it is very difficult to reduce the error floor to a requirement of a system by using the foregoing solutions, and an error correction method during decoding is complicated.